sheikdawoodraj...@gmail.com wrote:
[]
> Steve,
> what do you exactly mean "ultra high quality time"?
>
> My ntp service is started running the command below:
> "C:\ntp\bin\ntpd.exe" -U 3 -M -g -c "C:\ntp\etc\ntp.conf"
>
> Are you talking about the -M (Windows only - set high quality
> multimedia ti
On Apr 6, 4:44 pm, mi...@udel.edu (David Mills) wrote:
> Steve,
>
> You scratch a sore spot as I verified over the last couple of days. You
> can always increase the mindist, but eventually the orphan servers will
> drift away from each other. The problem is the orphan server has to be
> selected b
On Apr 7, 5:18 am, Steve Kostecke wrote:
> On 2009-04-07, David Woolley wrote:
>
> > Steve Kostecke wrote:
>
> >> While ordinary NTP clients use a selection metric based on delay and
> >> dispersion, orphan children use a metric computed from the IP address of
> >> each core server. Each orphan c
On Apr 6, 7:05 pm, Steve Kostecke wrote:
> On 2009-04-06, sheikdawoodraj...@gmail.com
> wrote:
>
> [snip: a 27 line block of text]
>
> > Yes, that's what I'm experimenting now.
>
> What are you experimenting with?
>
> > I beleive the Orphan Broadcast Mesh will work when none of the
> > systems c
On Apr 6, 6:35 pm, Steve Kostecke wrote:
> On 2009-04-07, sheikdawoodraj...@gmail.com
> wrote:
>
> > On Apr 6, 12:32 pm, Steve Kostecke wrote:
>
> >> OrphanMode is intended to select _one_ leader from the group of
> >> Orphans when all other time sources are unreachable. It does so based
> >> on
On Apr 6, 4:44 pm, mi...@udel.edu (David Mills) wrote:
> Steve,
>
> You scratch a sore spot as I verified over the last couple of days. You
> can always increase the mindist, but eventually the orphan servers will
> drift away from each other. The problem is the orphan server has to be
> selected b
On Apr 6, 7:05 pm, Steve Kostecke wrote:
> On 2009-04-06, sheikdawoodraj...@gmail.com
> wrote:
>
> [snip: a 27 line block of text]
>
> > Yes, that's what I'm experimenting now.
>
> What are you experimenting with?
>
> > I beleive the Orphan Broadcast Mesh will work when none of the
> > systems c
On Apr 6, 6:53 pm, Steve Kostecke wrote:
> On 2009-04-06, David Woolley wrote:
>
> > Steve Kostecke wrote:
>
> >> "Sync" is usually defined as "calculating the offset between a clock and
> >> a reference standard, stepping or initiating a slew to correct the
>
> > I doubt that many would consider
On Apr 6, 7:05 pm, Steve Kostecke wrote:
> On 2009-04-06, sheikdawoodraj...@gmail.com
> wrote:
>
> [snip: a 27 line block of text]
>
> > Yes, that's what I'm experimenting now.
>
> What are you experimenting with?
>
> > I beleive the Orphan Broadcast Mesh will work when none of the
> > systems c
Steve Kostecke wrote:
> On 2009-04-07, David Woolley wrote:
>
>> Steve Kostecke wrote:
>>
>>> It is imcumbent on the operator of an ntpd (or some "ntp client") to
>>> use enough time sources so that an outlier that is off by 120ms will
>>> be detected and voted out.
>> Unruh actually meant 128 -
Dave,
The leap bits are set the first time a valid update is received and the
clock is set. If the NMEA driver provides thethat update, it will set
the server leap bits. The leap value 11 is set only before first
synchronized and never set back to that value after that. Whether or not
in FREQ
Someone has observed that they get different behavior on 4.2.5p161
regarding the leap bits during freq_mode depending on whether they
configure a refclock (NMEA in this case) or not. With network
upstream servers configured, they observe leap=00 shortly after
starting, during the frequency drift c
Steve,
I've tested the time island before and it works even for long periods,
because once the children have all agreed on a single parent, they
dutifully follow the parent no matter what it does. My latest test was
different; it involved two distinct and separate orphan parents, each
drifting
Steve Kostecke wrote:
> On 2009-04-06, David Woolley wrote:
>> Steve Kostecke wrote:
>>
>>> "Sync" is usually defined as "calculating the offset between a clock and
>>> a reference standard, stepping or initiating a slew to correct the
>>
>> I doubt that many would consider the initiation of a sle
Hi,
I have 4 servers that should be in sync for time with each other as well
as much as the upstream ntp servers. Am I right that I can configure
this like this or is it that if you configure peers the upstream servers
are ignored?
# peer a <- this system is a
peer b
peer c
peer d
server ntp.nmi
On 2009-04-07, David Woolley wrote:
> Steve Kostecke wrote:
>
>> While ordinary NTP clients use a selection metric based on delay and
>> dispersion, orphan children use a metric computed from the IP address of
>> each core server. Each orphan child chooses the orphan parent as the
>> root server w
On 2009-04-07, David Woolley wrote:
> Steve Kostecke wrote:
>
>> It is imcumbent on the operator of an ntpd (or some "ntp client") to
>> use enough time sources so that an outlier that is off by 120ms will
>> be detected and voted out.
>
> Unruh actually meant 128 - epsilon milliseconds. The outl
On 2009-04-07, David Woolley wrote:
> Danny Mayer wrote:
>
>> I know that. -o is specific to windows AFAIK.
>
> It's specific to some/all versions of NT. It does exist on Linux, but
> seems to do something different.
According to the netstat man page on one of my Debian systems:
-o, --timers
David J Taylor wrote:
> David Woolley wrote:
>> David J Taylor wrote:
>>
>>> netstat -ano works correctly on my Windows XP Pro SP3.
>>>
>> I ran strings on the Win 98 version to find the help message, so it
>> looks like it has changed on XP, although ntpd wouldn't run on 98,
>> anyway.
>
> Checke
David Woolley wrote:
> David J Taylor wrote:
>
>> netstat -ano works correctly on my Windows XP Pro SP3.
>>
>
> I ran strings on the Win 98 version to find the help message, so it
> looks like it has changed on XP, although ntpd wouldn't run on 98,
> anyway.
Checked on my Windows 2000 system and t
Danny Mayer wrote:
>
> I know that. -o is specific to windows AFAIK.
It's specific to some/all versions of NT. It does exist on Linux, but
seems to do something different.
___
questions mailing list
questions@lists.ntp.org
https://lists.ntp.org/mail
David J Taylor wrote:
> netstat -ano works correctly on my Windows XP Pro SP3.
>
I ran strings on the Win 98 version to find the help message, so it
looks like it has changed on XP, although ntpd wouldn't run on 98, anyway.
___
questions mailing list
Steve Kostecke wrote:
> On 2009-04-06, sheikdawoodraj...@gmail.com
> wrote:
> >
> I've conducted extensive testing with Orphan Mode and never had to add
> a tos mindist command.
He has that because Dr Mills has just told him he needs it.
>
>> I beleive the "time island" concept will work fine
sheikdawoodraj...@gmail.com wrote:
>
> [Dawood] Does any ntp variable holds the value of "sync distance" ?
> something like "rootdelay" and "rootdispersion" variables.
It's a combination of various things, of which the most important are
0.5* rootdelay, 0.5*last hop delay, rootdispersion, local
Steve Kostecke wrote:
>
> Most users, though, consider their clocks to be "synced" when they see
> the asterisk indicating that ntpd has selected a sys_peer.
>
That's because they are making the, unsafe, assumption that ntpd is
using a reasonable definition of synched.
Note that my definition
Steve Kostecke wrote:
> It is imcumbent on the operator of an ntpd (or some "ntp client") to
> use enough time sources so that an outlier that is off by 120ms will be
> detected and voted out.
Unruh actually meant 128 - epsilon milliseconds. The outlyer is the
machine that is just starting!
If
Steve Kostecke wrote:
>
> While ordinary NTP clients use a selection metric based on delay and
> dispersion, orphan children use a metric computed from the IP address of
> each core server. Each orphan child chooses the orphan parent as the
> root server with the smallest metric.
From what Dr M
27 matches
Mail list logo